Subwoofer / Low Frequency Optimization

Hi guys. For the last 3-4 years I have been really keen on unlocking the mystery of room of acoustics. I read every forum post I could, white papers, you name it. None of them helped as there is so much conflicting advice out there.

In finally found one consistent story that fits completely together. And that is the teachings of Dr. Toole of Harman/NRC fame. Through lots of face time, classes, reading of his papers (and that of other Harman researchers), seat time at their labs, and our own theater design, combined with quality time with Keith Yates, it all became a compelling story that works at theory and practical point of views.

Instead of having you all follow the same journey, I thought I summarize my learning. What you are about to read was slated to be published by Widescreen Review Magazine but I missed the deadline and I decided to put it online. The topic is as this thread says: how to get high-fidelity bass reproduction in a theater or listening space. http://www.madronadigital.com/Library/BassOptimization.html

So allocate 15 to 20 minutes and read through it. I think you will find it highly worthwhile.

Happy to field questions about it and expand on any topic there.

Further installments will focus on the rest of the story above low frequencies.

So let me know what you think. It can be anything from spelling errors to "you have no idea what you are talking about."

Thanks for all the comments guys. I like to make sure we don't think the article or the solution is strictly about multiple subwoofers. It is not. The message is broader than that:

1. When people think of dealing with the room, they immediately think of putting panels on the walls. That is wrong on multiple fronts. I will address that in a future article. But for now, the case is clearly made in the article with measurements that the impact of the room is the most in lower frequencies. In the higher frequencies the speaker is in control. A good speaker in a typical room may not need any treatment to produce great sound. But the same great speaker will have totally non-flat response in low frequencies.

2. Typical acoustic treatment panels is lucky to go down to the transition frequency of a few hundred hertz. It has no impact in the "modal region" (e.g. < 100 Hz). So if that is all you have on your walls, you have not dealt with low frequencies.

3. Unless you go to extreme measures (full CFD analysis, professionally designed room and multiple subs) you are not going to have anything close to a flat response. An EQ is *mandatory* to get smooth frequency response. Even our $300K reference theater needs it as the ARCOS measurement shows. You want to use multiple subs to get rid of nasty nulls (subtractions) and then whatever peaks that are left, fix with an EQ. The EQ does not need to be automatic but it must be there if you care about solving this problem.

Remember, without it you can easily have a 10db bump that will swamp the clarity of the rest of the rest of the frequencies. Part of the reason for this is that the behavior in frequency domain like that, translates to "ringing" in time domain. If the ringing is higher than about 0.4 sec which is the rhythm in a lot of music, you muddy up what you are hearing. Bring down the frequency response hump and the ringing is reduced with it. You can see this in action in Dr. Toole's book:



See how reducing the peak on the left in frequency response, resulted in the impulse lasting less in time domain and ringing reduced (one is for crude EQ and the other, fine resolution).

4. On Earl's method, I have read all the threads and at the end, there is no process there other than randomly placing "subs" out there. Todd's paper actually simulates random placement and shows that it is not better. The ideal solution is CFD modelling but if you have a rectangular room, going by analyzed positions is the best route. Note that more subs is NOT always better. The subs are fighting each other if you will and your output can sharply go down in many of the scenarios. Todd simulates 5,000 subs and from what I recall, it had lower output than just using two! Indeed, the two sub method has more output than 4. Four subs in the centers gets you smoother response than four subs in the corner but the latter has a lot more power. So with four, the recommended solution is actually to put them in the corners as a better trade off. In our secondary sub configuration we have them in the center points but we have doubled up the subs (four drivers each) to get enough bass.

kevinh said:

Very good article Amir, lays out a systematic process for getting good base in a room.

The JBL equipment sounds very interesting, would be interested in knowing how different
it is from the Tact and DeqX in handling low freq. The CFD software also looks intriguing,
giving one the ability to look at subs at different heights as well at xy positioning in a roomprior to
physically measuring and making adjustments.

Great point about optimizing prior to applying room eq also.

Click to expand...

As an owner of TacT, I can comment on that. I have only heard Deqx at a show but no seat time with it. Here are the differences:

Positives for JBL:
1. It optimizes for up to 8 seats. TacT only works for one. So for theater applications or wanting to have more than one seat with good sound, TacT does not work.

2. It works! The hardware for the SDEC-4500 is actually from their professional BSS division. It is a general purpose EQ made to be beat to death in live concert situations and such. It is industrial quality and designed to last. My TacT bit the dust and is not supported by the company. The difference here could not be greater.

3. The process is real-time. You can grab any filter point and move it and hear its effect as the music is playing. TacT is a batch operation and doesn't let you change the automatic EQ that way. At least I don't recall it doing so.

4. JBL has built-in support for up to 4 subs and uses it for Sound Field Management to get smoother response by varying what each sub is playing. TacT doesn't have this logic (Harman has patented this).

5. The JBL system is the only one that has passed a double blind test.

Negatives for JBL:
1. It is dealer-only system. The measurement kit is separate from the box that does the work. The dealer brings the kit, calibrates your system and then the system runs stand-alone. The kit hardware is quite high-end, using a sophisticated DSP of its own.

For users who want to continue to play with their system per above, we can enable that in that we can give you full control to all the parameters created by the measurement kit. More clearly, all the measurement kit does is to program the DSP on what to do. That program is left with the system and is the thing that does the optimization. You can modify that program and it is pretty easy since all it is doing is applying filters, levels and timing.

2. It is not cheap. The retail for it is $16K. But for a system with so many inputs and outputs, it is pretty reasonable I would say . Currently JBL has "b-stock" units that are ridiculously cheap. From what I have heard, they are as good as new as they heavily refurbish them. So if you are tempted by this technology, contact your JBL Synthesis dealer and inquire about this.

3. The unit has pretty loud fans and the normal placement is outside of listening room. The fans are however there to deal with the rough commercial sound reinforcement where the conditions can be quite hostile. In typical applications the unit is not working hard at all and we can disable the fan on request.

4. The box is analog in, analog out. It is possible to get digital I/O for it but in standard mode, it re-digitizes the sound. In my view, the improvements it makes outweigh any drawback here. Our ears are far more sensitive to linear distortion such as room induced frequency response variations than non-linear distortion created by the ADC/DAC.

I just bought this unit for myself . Even at my cost it is still is significant outlay. But as I post above, there is really no way to get proper sound without an EQ. And once you hear proper EQ, you don't want to do without it.

Thanks Steve. The audio portion of the project is still a going concern but not definite plans as what may come from it.

For now, one building block is there in that Harman is now the sole distributor for Lumagen video processor. The rumor is that this can be augmented with an-audio-only processor to create a complete AV solution from Harman/Lexicon. I can't tell if that is going to be the case but it is a step to the right direction .

My current solution is to a $1,200 Marantz AV processor with balanced out to SDEC-4500, then out to the speakers. The Marantz is rock solid as far as video/audio switching. And the SDEC then provides the real audio processing.